library(magrittr)
library(tidyverse)
library(Seurat)
library(readxl)
library(cowplot)
library(colorblindr)
library(viridis)
library(progeny)
library(destiny)

theme_cowplot2 <- function(...) {
  theme_cowplot(font_size = 12, ...) %+replace%
    theme(strip.background = element_blank(),
          plot.background = element_blank())
}
theme_set(theme_cowplot2())
coi <- params$cell_type_super
cell_sort <- params$cell_sort
cell_type_major <- params$cell_type_major
louvain_resolution <- params$louvain_resolution
louvain_cluster <- params$louvain_cluster

1 Cluster markers

1.1 Major T.super markers for cell assign

### load all data ---------------------------------
source("_src/global_vars.R")

# seu_obj <- read_rds(paste0("/work/shah/isabl_data_lake/analyses/16/52/1652/celltypes/", coi, "_processed.rds"))
seu_obj <- read_rds(paste0("/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/outs_prepost/", coi, "_seurat_", louvain_resolution, ".rds"))

myfeatures <- c("umapharmony_1", "umapharmony_2", "sample", louvain_cluster, "doublet", "nCount_RNA", "nFeature_RNA", "percent.mt", "doublet_score")

plot_data_wrapper <- function(cluster_res) {
  cluster_res <- enquo(cluster_res)
  as_tibble(FetchData(seu_obj, myfeatures)) %>% 
    left_join(meta_tbl, by = "sample") %>% 
    rename(cluster = !!cluster_res) %>% 
    mutate(cluster = as.character(cluster),
           tumor_supersite = ordered(tumor_supersite, levels = names(clrs$tumor_supersite)))
}

plot_data <- plot_data_wrapper(louvain_cluster)

1.2 Subtype currated markers

helper_f <- function(x) ifelse(is.na(x), "", x)

markers_v6_super[[coi]] %>% 
  group_by(subtype) %>% 
  mutate(rank = row_number(gene)) %>% 
  spread(subtype, gene) %>% 
  mutate_all(.funs = helper_f) %>% 
  formattable::formattable()

rank	CD4.T.dysfunctional	CD4.T.naive	CD4.T.reg	CD8.T	Cycling.T.NK	MT.high.T.NK	NK.CD56	NK.Cytotoxic
1	CD4	CCR7	CD4	CCL4	CDC20	IGKC	CD63	ADGRG1
2	CD40LG	CD4	FOXP3	CD8A	CDK1	MALAT1	CD7	CX3CR1
3	CTLA4	IL7R	IL2RA	CD8B	MKI67	MIAT	FCER1G	FCER1G
4	CXCL13	KLF2	TNFRSF4	CRTAM	PTTG1	MT-ND6	GNLY	FCGR3A
5	FKBP5	TCF7	TRAC	GZMA	TOP2A	MTRNR2L12	KLRC1	FGFBP2
6	IL6ST			GZMB		XIST	KLRD1	GNLY
7	ITM2A			GZMK			KLRF1	GZMH
8	MAF			GZMM			KRT81	IGFBP7
9	NMB			IFNG			KRT86	KLRD1
10	NR3C1			LAG3			NCAM1	KLRF1
11	PDCD1			MT1E			NKG7	NKG7
12	TNFRSF4			MT1X			TYROBP	PLAC8
13	TOX2			MT2A			XCL1	PLEK
14	TSHZ2			PTMS			XCL2	PTGDS
15				TRGC2				SPON2
16								TYROBP

1.3 Subtype cluster markers

# marker_tbl <- read_tsv(paste0("/work/shah/isabl_data_lake/analyses/16/52/1652/celltypes/", coi, "_markers.tsv")) %>% 
#   filter(resolution == louvain_resolution)
marker_tbl <- read_tsv(paste0("/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/outs_prepost/", coi, "_markers_", louvain_resolution, ".tsv"))

## Hypergeometric test --------------------------------------

test_set <- marker_tbl %>% 
  group_by(cluster) %>% 
  filter(!str_detect(gene, "^RPS|^RPL")) %>% 
  slice(1:50) %>% 
  mutate(k = length(cluster)) %>% 
  ungroup %>%
  select(cluster, gene, k) %>% 
  mutate(join_helper = 1) %>% 
  group_by(cluster, join_helper, k) %>%
  nest(test_set = gene)

markers_doub_tbl <- markers_v6 %>% 
  enframe("subtype", "gene") %>% 
  filter(!(subtype %in% unique(c(coi, cell_type_major)))) %>% 
  unnest(gene) %>% 
  group_by(gene) %>% 
  filter(length(gene) == 1) %>% 
  mutate(subtype = paste0("doublet.", subtype)) %>% 
  bind_rows(tibble(subtype = "Mito.high", gene = grep("^MT-", rownames(seu_obj), value = T)))

ref_set <- markers_v6_super[[coi]] %>% 
  bind_rows(markers_doub_tbl) %>% 
  group_by(subtype) %>% 
  mutate(m = length(gene),
         n = length(rownames(seu_obj))-m,
         join_helper = 1) %>% 
  group_by(subtype, m, n, join_helper) %>%
  nest(ref_set = gene)

hyper_tbl <- test_set %>% 
  left_join(ref_set, by = "join_helper") %>% 
  group_by(cluster, subtype, m, n, k) %>%
  do(q = length(intersect(unlist(.$ref_set), unlist(.$test_set)))) %>%
  mutate(pval = 1-phyper(q = q, m = m, n = n, k = k)) %>%
  ungroup %>%
  mutate(qval = p.adjust(pval, "BH"),
         sig = qval < 0.01)

# hyper_tbl %>% 
#   group_by(subtype) %>% 
#   filter(any(qval < 0.01)) %>%
#   ggplot(aes(subtype, -log10(qval), fill = sig)) +
#   geom_bar(stat = "identity") +
#   facet_wrap(~cluster) +
#   coord_flip()
  
low_rank <- str_detect(unique(hyper_tbl$subtype), "doublet")
subtype_lvl <- c(sort(unique(hyper_tbl$subtype)[!low_rank]), sort(unique(hyper_tbl$subtype)[low_rank]))
  
cluster_label_tbl <- hyper_tbl %>% 
  mutate(subtype = ordered(subtype, levels = subtype_lvl)) %>% 
  arrange(qval, subtype) %>%
  group_by(cluster) %>% 
  slice(1) %>% 
  mutate(subtype = ifelse(sig, as.character(subtype), paste0("unknown_", cluster))) %>% 
  select(cluster, cluster_label = subtype) %>% 
  ungroup %>% 
  mutate(cluster_label = make.unique(cluster_label, sep = "_"))

seu_obj$cluster_label <- unname(deframe(cluster_label_tbl)[as.character(unlist(seu_obj[[paste0("RNA_snn_res.", louvain_resolution)]]))])
plot_data$cluster_label <- seu_obj$cluster_label

marker_sheet <- marker_tbl %>% 
  left_join(cluster_label_tbl, by = "cluster") %>% 
  group_by(cluster_label) %>% 
  filter(!str_detect(gene, "^RPS|^RPL")) %>% 
  slice(1:50) %>% 
  mutate(rank = row_number(-avg_logFC)) %>% 
  select(cluster_label, gene, rank) %>% 
  spread(cluster_label, gene) %>% 
  mutate_all(.funs = helper_f)

formattable::formattable(marker_sheet)

rank	CD4.T.dysfunctional	CD4.T.naive	CD4.T.reg	CD8.T	CD8.T_1	Cycling.T.NK	doublet.B.cell	doublet.Fibroblast	doublet.Monocyte	MT.high.T.NK	NK.CD56	NK.Cytotoxic	unknown_12
1	CXCL13	IL7R	TNFRSF4	CD8A	APOE	STMN1	IGHM	DCN	CST3	IGKC	KLRC1	FGFBP2	LST1
2	NMB	CCR7	FOXP3	CD8B	HLA-DRB5	MKI67	MS4A1	IGFBP5	S100A8	MALAT1	TYROBP	FCGR3A	AREG
3	MAF	CD40LG	IL2RA	GZMK	ATP5F1E	TOP2A	MEF2C	RBP1	S100A9	MIAT	GNLY	SPON2	IL4I1
4	TSHZ2	MAL	RTKN2	CCL4L2	CCL3L1	CENPF	CD79A	C7	CXCL8	IGLC3	FCER1G	KLRF1	PCDH9
5	IL6ST	TCF7	BATF	CRTAM	GZMK	ASPM	BANK1	CALD1	SPP1	NEAT1	AREG	CX3CR1	SCN1B
6	CD40LG	GPR183	CTLA4	TRGC2	SH2D1A	TYMS	BCL11A	MEG3	HLA-DRA	IGLC2	TRDC	PRF1	SPINK2
7	CPM	EEF1B2	TBC1D4	GZMH	C1QB	TUBA1B	TNFRSF13C	SOX4	LYZ	NABP1	KRT81	PLAC8	TNFSF13B
8	ZBED2	LEF1	GADD45A	LAG3	SPP1	NUSAP1	MARCH1	RARRES2	APOE	SPOCK2	XCL1	PTGDS	KIT
9	IGFL2	TPT1	LTB	IFNG	PRMT9	TUBB	VPREB3	IGFBP4	C1QB	IL7R	IGFBP2	PLEK	IL1R1
10	PDCD1	SELL	TNFRSF18	CCL5	HIST1H4C	PCLAF	RALGPS2	NR2F2	AIF1	CD84	XCL2	GNLY	LTC4S
11	CD4	EEF1A1	PMAIP1	CCL4	CD8B	HMGB2	IGHD	ADIRF	MARCKS	CD2	KRT86	KLRD1	TOX2
12	ICA1	AQP3	IKZF2	ITM2C	APOC1	HIST1H4C	TCF4	SELENOP	C1QA	CLEC2D	KLRD1	ADGRG1	CSF2
13	TOX2	TRABD2A	SOX4	DTHD1	IL32	HIST1H1B	BASP1	MDK	FN1	CD3G	CEBPD	CLIC3	KRT81
14	CTLA4	LTB	SAT1	MT1X	GPR183	TPX2	ARHGAP24	MGP	FTL	B2M	CLIC3	IGFBP7	AFF3
15	TNFRSF4	ANK3	TIGIT	JAML	SLC2A3	UBE2C	CD83	TAGLN	APOC1	CD3D	TXK	NKG7	SCX
16	ITM2A	CTSL	ICOS	CCL3L1	KLRG1	CLSPN	CD22	STAR	G0S2	HLA-A	TMIGD2	GZMB	PLAT
17	CD200	TNFRSF25	TNFRSF9	TNIP3	CKAP2	PCNA	SWAP70	ADAMTS1	MNDA	CD74	MATK	PRSS23	IL23R
18	PTPN13	PASK	LAYN	CXCR6	CD3G	HMGN2	ADAM28	LUM	BASP1	CYBA	IL2RB	AKR1C3	TLE1
19	CCDC50	DPP4	TNFRSF1B	TNFSF9	CD8A	SMC4	AFF3	SFRP4	MS4A6A	HLA-B	KLRB1	EFHD2	SVIL
20	CHN1	ITGA6	F5	THEMIS	CCL4L2	KNL1	CYBB	C1R	IL1B	HSPH1	KLRC2	S1PR5	CD300LF
21	BHLHE40-AS1	CMTM8	IL1R1	HLA-DPB1	BCL2A1	SMC2	TCL1A	PEG3	CD14	HLA-C	NCAM1	TYROBP	RORC
22	CH25H	MYC	IL32	MT1E	HMGA1	H2AFZ	CD24	SPARCL1	C1QC	MAT2A	CTSW	FCER1G	TNFSF11
23	SMCO4	TSHZ2	CD4	HLA-DRB1	CD2	CKS1B	HVCN1	FHL2	LST1	FOSB	HOPX	GZMH	LIF
24	BTLA	TOB1	MIR4435-2HG	GZMA	PTGER4	BIRC5	BLK	C1S	CSF3R	RCSD1	LAT2	CST7	TGM2
25	DUSP4	RACK1	IL1R2	KLRG1	PECAM1	CENPE	LY86	AKAP12	FTH1	SPTBN1	CXXC5	HOPX	ZG16B
26	GNG4	CCR6	ENTPD1	HLA-DPA1	GIMAP7	RRM2	FCRL5	NR2F1	GRN	CXCR3	CD7	MYOM2	ENPP1
27	SESN3	CYSLTR1	ICA1	DUSP2	GZMA	DLGAP5	FCRL1	SERPINF1	CYBB	ERO1A	SH2D1B	MYBL1	IRAK3
28	CYSLTR1	FXYD1	BIRC3	MT2A	CST7	DUT	SCIMP	CLU	NPC2	TXK	SRGAP3	CHST2	PTGER3
29	BATF	LPAR6	IFI27	COTL1		ATAD2	PAX5	EGR1	SPI1	GRAP2	CCL3	CEP78	FOXC1
30	MIR155HG	CHD7	CD27	APOBEC3G		CDK1	BLNK	WFDC2	RNASE1	RUNX1	TNFRSF18	CCL3	CA2
31	RILPL2	UBA52	GLRX	CST7		HELLS	SPIB	SERPING1	FCGR2A	TCF7	CD63	CD247	KIAA1211L
32	TNFRSF25	JUNB	TYMP	FABP5		GTSE1	PKIG	PTCRA	SERPINA1	ITGB7	GSTP1	MTSS1	COL4A4
33	ETV7	TIMP1	SLAMF1	PDCD1		CDKN3	CD19	CDKN1C	PLAUR	POU2F2	NKG7	CD300A	SLC4A10
34	TNFRSF18	SESN3	BTG3	CD3D		UBE2S	GNG7	TCEAL9	GLUL	VSIR	KLRF1	FGR	NRP1
35	CTSL	NACA	DUSP4	MT1F		ZWINT	COBLL1	TPM2	MS4A7	HAVCR2	MCTP2	PTGDR	PDZK1
36	MYO7A	CD4	PIM2	CD3G		CKS2	FAM30A	DLK1	CD83	LAT	IFITM3	GSAP	APOL4
37	CD84	RASGRP2	LAIR2	ZNF683		PTTG1	CD40	IFITM3	MAFB	SRGN	ITGAX	SH2D1B	B3GALT5
38	ENTPD1	FBLN5	MAF	CCR5		KIF11	WDFY4	COL6A1	IER3	MYO1G	KIR2DL4	ADRB2	BCAS1
39	TP53INP1	SATB1	MAST4	EOMES		H2AFX	FCRL2	COL1A2	CD68	PPT1	ZNF683	CTSW	PRAM1
40	ADAM19	S1PR1	HPGD	GZMM		TK1	BTK	FILIP1L	CPVL	ZEB1	MAFF	FCRL6	NECTIN2
41	ICOS	CAMK4	S100A4	CD27		HMMR	TLR10	NUPR1	ALDH2	TIMP1	SLC16A3	XBP1	LINGO4
42	RGS1	RIPOR2	SPOCK2	OASL		MCM7	GAPT	SLC40A1	CSF2RA	ICAM2	ATP8B4	TXK	HOXA5
43	SRGN	ANXA1	PHLDA1	DNAJB1		TMPO	POU2AF1	MARCKSL1	FGL2	GPR174	B3GNT7	LYN	PRKG1
44	CD82	MGAT4A	LAPTM4B	PECAM1		TUBB4B	EBF1	TSC22D1	OLR1	NAMPT	ADGRG3	TGFBR3	PDZD2
45	TOX	TMEM204	HERC5	MIR155HG		BRCA1	FCRLA	CST3	GPNMB	KLRG1	GZMB	CD160	SLCO2A1
46	SPOCK2		LTA	SIT1		EZH2	CPNE5	RARRES1	FPR1	KDM6B	ITGA1	KLRB1	PKDCC
47	PRDM1		ETV7	GPR174		MCM4	OSBPL10	GNG11	SLC11A1	TMEM156	ZBTB16	AREG	DACH1
48	DGKH		POU2F2	GRAP2		MAD2L1	FAM129C	CFH	TGFBI	SESN3	CLNK	GPR65	BICC1
49	BIRC3		NCF4	IKZF3		CENPM	HLA-DOB	PDGFRA	TNFAIP2	FKBP11	ADAM28	ITGAM	EMID1
50	TRAT1		ID3	ITGA1		CEP55	SSPN	FBLN1	CCDC88A	GALM	CD160	SAMD3	CD33

write_tsv(marker_sheet, paste0("/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/", coi, "_marker_sheet.tsv"))

2 Clusters

2.1 sizes

enframe(sort(table(seu_obj$cluster_label))) %>% 
  mutate(name = ordered(name, levels = rev(name))) %>% 
  ggplot() +
  geom_bar(aes(name, value), stat = "identity") +
  theme(axis.text.x = element_text(angle = 90, hjust = 1, vjust = 0.5)) +
  labs(y = c("#cells"), x = "cluster")

2.2 UMAP

alpha_lvl <- ifelse(nrow(plot_data) < 20000, 0.2, 0.1)
pt_size <- ifelse(nrow(plot_data) < 20000, 0.2, 0.05)

common_layers_disc <- list(  
  geom_point(size = pt_size, alpha = alpha_lvl),
  NoAxes(),
  guides(color = guide_legend(override.aes = list(size = 2, alpha = 1))),
  labs(color = "")
)

common_layers_cont <- list(  
  geom_point(size = pt_size, alpha = alpha_lvl),
  NoAxes(),
  scale_color_gradientn(colors = viridis(9)),
  guides(color = guide_colorbar())
)

ggplot(plot_data, aes(umapharmony_1, umapharmony_2, color = cluster_label)) + 
  common_layers_disc +
  #facet_wrap(~therapy) +
  ggtitle("Sub cluster")

3 Filtering out doublet and mito high clusters

3.1 UMAP

my_subtypes <- names(clrs$cluster_label[[coi]])

my_subtypes <- c(my_subtypes, unlist(lapply(paste0("_", 1:3), function(x) paste0(my_subtypes, x)))) %>% .[!str_detect(., "doublet")]

cells_to_keep <- colnames(seu_obj)[seu_obj$cluster_label %in% my_subtypes]
# seu_obj_sub <- subset(seu_obj, cells = cells_to_keep)
# seu_obj_sub <- RunUMAP(seu_obj_sub, dims = 1:50, reduction = "harmony", reduction.name = "umapharmony", reduction.key = "umapharmony_")
# seu_obj_sub$cluster_label <- seu_obj$cluster_label[colnames(seu_obj) %in% colnames(seu_obj_sub)]
# write_rds(seu_obj_sub, paste0("/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/", coi, "_processed_filtered.rds"))
seu_obj_sub <- read_rds(paste0("/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/", coi, "_processed_filtered.rds"))

plot_data_sub <- as_tibble(FetchData(seu_obj_sub, c(myfeatures, "cluster_label"))) %>% 
  left_join(meta_tbl, by = "sample") %>% 
  mutate(patient_id = str_remove_all(patient_id, "SPECTRUM-OV-"),
         tumor_supersite = ordered(tumor_supersite, levels = names(clrs$tumor_supersite))) %>% 
  mutate(cell_id = colnames(seu_obj_sub),
         cluster_label = ordered(cluster_label, levels = my_subtypes),
         )
  
if (cell_sort == "CD45+") {
plot_data_sub <- filter(plot_data_sub, sort_parameters != "singlet, live, CD45-", !is.na(tumor_supersite))
}

if (cell_sort == "CD45-") {
plot_data_sub <- filter(plot_data_sub, sort_parameters != "singlet, live, CD45+", !is.na(tumor_supersite))
}

ggplot(plot_data_sub, aes(umapharmony_1, umapharmony_2, color = cluster_label)) + 
  common_layers_disc +
  ggtitle("Sub cluster") +
  #facet_wrap(~cluster_label) +
  scale_color_manual(values = clrs$cluster_label[[coi]])

ggplot(plot_data_sub, aes(umapharmony_1, umapharmony_2, color = patient_id_short)) + 
  common_layers_disc +
  ggtitle("Patient") +
  # facet_wrap(~therapy) +
  scale_color_manual(values = clrs$patient_id_short)

ggplot(plot_data_sub, aes(umapharmony_1, umapharmony_2, color = tumor_supersite)) + 
  # geom_point(aes(umapharmony_1, umapharmony_2), 
  #            color = "grey90", size = 0.01, 
  #            data = select(plot_data_sub, -tumor_supersite)) +
  common_layers_disc +
  ggtitle("Site") +
  # facet_wrap(~therapy) +
  scale_color_manual(values = clrs$tumor_supersite)

write_tsv(select(plot_data_sub, cell_id, everything(), -umapharmony_1, -umapharmony_2, -contains("RNA_")), paste0("/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/", coi, "_embedding.tsv"))

3.2 add module scores and pathway scores

signature_modules <- read_excel("_data/small/signatures/SPECTRUM v5 sub cluster markers.xlsx", sheet = 2, skip = 1, range = "M2:P100") %>% 
  gather(module, gene) %>% 
  na.omit() %>% 
  group_by(module) %>% 
  do(gene = c(.$gene)) %>% 
  {setNames(.$gene, .$module)}

signature_modules$ISG.module <- c("CCL5", "CXCL10", "IFNA1", "IFNB1", "ISG15", "IFI27L2", "SAMD9L")

## compute expression module scores
for (i in 1:length(signature_modules)) {
  seu_obj_sub <- AddModuleScore(seu_obj_sub, features = signature_modules[i], name = names(signature_modules)[i])
  seu_obj_sub[[names(signature_modules)[i]]] <- seu_obj_sub[[paste0(names(signature_modules)[i], "1")]]
  seu_obj_sub[[paste0(names(signature_modules)[i], "1")]] <- NULL
  print(paste(names(signature_modules)[i], "DONE"))
}

## [1] "CD8.Cytotoxic DONE"
## [1] "CD8.Dysfunctional DONE"
## [1] "CD8.Naive DONE"
## [1] "CD8.Predysfunctional DONE"
## [1] "ISG.module DONE"

## compute progeny scores
progeny_list <- seu_obj_sub@assays$RNA@data[VariableFeatures(seu_obj_sub),] %>% 
  as.matrix %>% 
  progeny %>% 
  as.data.frame %>% 
  as.list

names(progeny_list) <- make.names(paste0(names(progeny_list), ".pathway"))

for (i in 1:length(progeny_list)) {
  seu_obj_sub <- AddMetaData(seu_obj_sub, metadata = progeny_list[[i]], 
                             col.name = names(progeny_list)[i])
}

write_rds(seu_obj_sub, paste0("/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/", coi, "_processed_filtered_annotated.rds"))

3.3 marker heatmap

marker_top_tbl <- marker_sheet[,-1] %>% 
  slice(1:10) %>% 
  as.list %>% 
  .[!str_detect(names(.), "doublet")] %>% 
  enframe("cluster_label_x", "gene") %>% 
  unnest(gene)

plot_data_markers <- as_tibble(FetchData(seu_obj_sub, c("cluster_label", myfeatures, unique(marker_top_tbl$gene)))) %>% 
  gather(gene, value, -c(1:(length(myfeatures)+1))) %>% 
  left_join(meta_tbl, by = "sample") %>% 
  mutate(patient_id = str_remove_all(patient_id, "SPECTRUM-OV-"),
         tumor_supersite = ordered(tumor_supersite, levels = names(clrs$tumor_supersite))) %>% 
  mutate(cluster_label = ordered(cluster_label, levels = my_subtypes)) %>% 
  group_by(cluster_label, gene) %>% 
  summarise(value = mean(value, na.rm = T)) %>% 
  group_by(gene) %>% 
  mutate(value = scales::rescale(value)) %>% 
  left_join(marker_top_tbl, by = "gene") %>% 
  mutate(cluster_label_x = ordered(cluster_label_x, levels = rev(names(clrs$cluster_label[[coi]]))))

ggplot(plot_data_markers) +
  geom_tile(aes(gene, cluster_label, fill = value)) +
  theme(axis.text.x = element_text(angle = 90, hjust = 1, vjust = 0.5)) +
  facet_grid(~cluster_label_x, scales = "free", space = "free") +
  scale_fill_gradientn(colors = viridis(9)) +
  labs(fill = "Scaled\nexpression") +
  theme(aspect.ratio = 1,
        axis.line = element_blank(),
        axis.ticks = element_blank(),
        axis.title = element_blank())

# ggsave(paste0("_fig/002_marker_heatmap_", coi, ".pdf"), width = nrow(marker_top_tbl)/6, height = 5)

3.4 composition

3.4.1 per site

comp_site_tbl <- plot_data_sub %>%
  filter(!is.na(tumor_supersite)) %>% 
  group_by(cluster_label, tumor_supersite) %>%
  tally %>%
  group_by(tumor_supersite) %>%
  mutate(nrel = n/sum(n)*100) %>%
  ungroup

pnrel_site <- ggplot(comp_site_tbl) +
  geom_bar(aes(tumor_supersite, nrel, fill = cluster_label),
           stat = "identity", position = position_stack()) +
  scale_y_continuous(expand = c(0, 0)) +
  theme(axis.text.x = element_text(angle = 90, hjust = 1, vjust = 0.5)) +
  labs(fill = "Cluster", y = "Fraction [%]", x = "") +
  scale_fill_manual(values = clrs$cluster_label[[coi]])

pnabs_site <- ggplot(comp_site_tbl) +
  geom_bar(aes(tumor_supersite, n, fill = cluster_label),
           stat = "identity", position = position_stack()) +
  scale_y_continuous(expand = c(0, 0)) +
  theme(axis.text.x = element_text(angle = 90, hjust = 1, vjust = 0.5)) +
  labs(fill = "Cluster", y = "# cells", x = "") +
  scale_fill_manual(values = clrs$cluster_label[[coi]])

plot_grid(pnabs_site, pnrel_site, ncol = 2, align = "h")

# ggsave(paste0("_fig/02_deep_dive_", coi, "_comp_site.pdf"), width = 8, height = 4)

3.4.2 per sample

comp_tbl_sample_sort <- plot_data_sub %>% 
  group_by(tumor_subsite, tumor_supersite, patient_id, consensus_signature, therapy, cluster_label) %>% 
  tally %>% 
  group_by(tumor_subsite, tumor_supersite, patient_id, consensus_signature, therapy) %>% 
  mutate(nrel = n/sum(n)*100,
         nsum = sum(n),
         log10n = log10(n),
         label_supersite = "Site",
         label_mutsig = "Signature",
         label_therapy = "Rx") %>% 
  ungroup %>% 
  arrange(desc(therapy), tumor_supersite) %>% 
  mutate(tumor_subsite_rx = paste0(tumor_subsite, "_", therapy)) %>% 
  mutate(tumor_subsite = ordered(tumor_subsite, levels = unique(tumor_subsite)),
         tumor_subsite_rx = ordered(tumor_subsite_rx, levels = unique(tumor_subsite_rx))) %>% 
  arrange(patient_id) %>% 
  mutate(label_patient_id = ifelse(as.logical(as.numeric(fct_inorder(as.character(patient_id)))%%2), "Patient1", "Patient2"))

sample_id_x_tbl <- plot_data_sub %>% 
  mutate(sort_short_x = cell_sort) %>% 
  distinct(patient_id, sort_short_x, tumor_subsite, therapy, sample) %>% 
  unite(sample_id_x, patient_id, sort_short_x, tumor_subsite, therapy) %>% 
  arrange(sample_id_x)

comp_tbl_sample_sort %>% 
  mutate(sort_short_x = cell_sort) %>% 
  unite(sample_id_x, patient_id, sort_short_x, tumor_subsite_rx) %>% 
  select(sample_id_x, cluster_label, n, nrel, nsum) %>% 
  left_join(sample_id_x_tbl, by = "sample_id_x") %>% 
  write_tsv(paste0("/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/", coi, "_subtype_compositions.tsv"))

ybreaks <- c(500, 1000, 2000, 4000, 6000, 8000, 10000, 15000, 20000)

max_cells_per_sample <- max(comp_tbl_sample_sort$nsum)
ymaxn <- ybreaks[max_cells_per_sample < ybreaks][1]

comp_plot_wrapper <- function(y = "nrel", switch = NULL) {
  if (y == "nrel") ylab <- paste0("Fraction\nof cells [%]")
  if (y == "n") ylab <- paste0("Number\nof cells")
  p <- ggplot(comp_tbl_sample_sort, 
              aes_string("tumor_subsite_rx", y, fill = "cluster_label")) + 
    facet_grid(~patient_id, space = "free", scales = "free", switch = switch) +
    coord_cartesian(clip = "off") + 
    scale_fill_manual(values = clrs$cluster_label[[coi]]) + 
    theme(axis.text.x = element_blank(),
          axis.title.y = element_text(angle = 0, vjust = 0.5, hjust = 0.5, 
                                      margin = margin(0, -0.4, 0, 0, unit = "npc")),
          axis.ticks.x = element_blank(),
          axis.title.x = element_blank(),
          axis.line.x = element_blank(),
          strip.text.y = element_blank(),
          strip.text.x = element_blank(),
          strip.background.y = element_blank(),
          strip.background.x = element_blank(),
          plot.margin = margin(0, 0, 0, 0, "npc")) + 
    labs(x = "",
         y = ylab) +
    guides(fill = FALSE)
  if (y == "nrel") p <- p + 
    geom_bar(stat = "identity") +
    scale_y_continuous(expand = c(0, 0), 
                       breaks = c(0, 50, 100), 
                       labels = c("0", "50", "100"))
  if (y == "n") p <- p + 
    geom_bar(stat = "identity", position = position_stack()) +
    scale_y_continuous(expand = c(0, 0), 
                       breaks = c(0, ymaxn/2, ymaxn),
                       limits = c(0, ymaxn),
                       labels = c("", ymaxn/2, ymaxn)) +
    expand_limits(y = c(0, ymaxn)) +
    theme(panel.grid.major.y = element_line(linetype = 1, color = "grey90", size = 0.5))
  return(p)
} 

common_label_layers <- list(
  geom_tile(color = "white", size = 0.15),
  theme(axis.text.x = element_blank(),
        axis.ticks = element_blank(),
        axis.title.x = element_blank(),
        axis.line.x = element_blank(),
        strip.text = element_blank(),
        strip.background = element_blank(),
        plot.margin = margin(0, 0, 0, 0, "npc")),
  scale_y_discrete(expand = c(0, 0)),
  labs(y = ""),
  guides(fill = FALSE),
  facet_grid(~patient_id, 
             space = "free", scales = "free")
)

comp_label_site <- ggplot(distinct(comp_tbl_sample_sort, tumor_subsite_rx, therapy, tumor_supersite, label_supersite, patient_id), 
       aes(tumor_subsite_rx, label_supersite, 
           fill = tumor_supersite)) + 
  scale_fill_manual(values = clrs$tumor_supersite) +
  common_label_layers

comp_label_rx <- ggplot(distinct(comp_tbl_sample_sort, tumor_subsite_rx, therapy, tumor_supersite, label_therapy, consensus_signature, patient_id), 
       aes(tumor_subsite_rx, label_therapy, 
           fill = therapy)) + 
  scale_fill_manual(values = c(`post-Rx` = "gold3", `pre-Rx` = "steelblue")) +
  common_label_layers

comp_label_mutsig <- ggplot(distinct(comp_tbl_sample_sort, tumor_subsite_rx, therapy, tumor_supersite, label_mutsig, consensus_signature, patient_id), 
       aes(tumor_subsite_rx, label_mutsig, 
           fill = consensus_signature)) + 
  scale_fill_manual(values = clrs$consensus_signature) +
  common_label_layers

patient_label_tbl <- distinct(comp_tbl_sample_sort, patient_id, .keep_all = T)

comp_label_patient_id <- ggplot(patient_label_tbl, aes(tumor_subsite_rx, label_patient_id)) + 
  scale_fill_manual(values = clrs$consensus_signature) +
  geom_text(aes(tumor_subsite_rx, label_patient_id, label = patient_id)) +
  facet_grid(~patient_id, 
             space = "free", scales = "free") +
  coord_cartesian(clip = "off") + 
  theme_void() +
  theme(strip.text = element_blank(),
        strip.background = element_blank(),
        plot.margin = margin(0, 0, 0, 0, "npc"))

hist_plot_wrapper <- function(x = "nrel") {
  if (x == "nrel") {
    xlab <- paste0("Fraction of cells [%]")
    bw <- 5
  }
  if (x == "log10n") {
    xlab <- paste0("Number of cells")
    bw <- 0.2
  }
  p <- ggplot(comp_tbl_sample_sort) +
    ggridges::geom_density_ridges(
      aes_string(x, "cluster_label", fill = "cluster_label"), color = "black",
      stat = "binline", binwidth = bw, scale = 3) +
    facet_grid(label_supersite~., 
               space = "free", scales = "free") +
    scale_fill_manual(values = clrs$cluster_label[[coi]]) + 
    theme(axis.text.y = element_blank(),
          axis.ticks.y = element_blank(),
          axis.title.y = element_blank(),
          axis.line.y = element_blank(),
          strip.text = element_blank(),
          strip.background = element_blank(),
          plot.margin = margin(0, 0, 0, 0, "npc")) +
    scale_x_continuous(expand = c(0, 0)) +
    scale_y_discrete(expand = c(0, 0)) +
    guides(fill = FALSE) +
    labs(x = xlab)
  if (x == "log10n") p <- p + expand_limits(x = c(0, 3)) + 
    scale_x_continuous(expand = c(0, 0), 
                       labels = function(x) parse(text = paste("10^", x)))
  return(p)
}

pcomp1 <- comp_plot_wrapper("n")
pcomp2 <- comp_plot_wrapper("nrel")

pcomp_grid <- plot_grid(comp_label_patient_id, 
                        pcomp1, pcomp2, 
                        comp_label_site, comp_label_rx, comp_label_mutsig,
                        ncol = 1, align = "v", 
                        rel_heights = c(0.15, 0.33, 0.33, 0.06, 0.06, 0.06))

phist1 <- hist_plot_wrapper("log10n")

pcomp_hist_grid <- ggdraw() +
  draw_plot(pcomp_grid, x = 0.01, y = 0, width = 0.85, height = 1) +
  draw_plot(phist1, x = 0.87, y = 0.05, width = 0.12, height = 0.8)

pcomp_hist_grid

# ggsave(paste0("_fig/02_composition_v6_",coi,".pdf"), pcomp_hist_grid, width = 10, height = 2)

3.4.3 site specific cluster enrichment

comp_tbl_z <- comp_tbl_sample_sort %>% 
  filter(therapy == "pre-Rx",
         !(tumor_supersite %in% c("Ascites", "Other"))) %>% 
  group_by(patient_id, cluster_label) %>% 
  arrange(patient_id, cluster_label, nrel) %>% 
  mutate(rank = row_number(nrel),
         z_rank = scales::rescale(rank)) %>% 
  mutate(mean_nrel = mean(nrel, na.rm = T),
         sd_nrel = sd(nrel, na.rm = T),
         z_nrel = (nrel - mean_nrel) / sd_nrel) %>% 
  ungroup()

ggplot(comp_tbl_z) +
  geom_boxplot(aes(tumor_supersite, z_nrel, color = tumor_supersite), 
               outlier.shape = NA) +
  geom_point(aes(tumor_supersite, z_nrel, color = tumor_supersite), position = position_jitter(), size = 0.1) +
  facet_grid(~cluster_label, scales = "free_x", space = "free_x") +
  theme(axis.text.x = element_text(angle = 90, hjust = 1, vjust = 0.5),
        aspect.ratio = 5) +
  scale_color_manual(values = clrs$tumor_supersite)

ggplot(comp_tbl_z) +
  geom_boxplot(aes(tumor_supersite, z_rank, color = tumor_supersite), 
               outlier.shape = NA) +
  geom_point(aes(tumor_supersite, z_rank, color = tumor_supersite), position = position_jitter(), size = 0.1) +
  facet_grid(~cluster_label, scales = "free_x", space = "free_x") +
  theme(axis.text.x = element_text(angle = 90, hjust = 1, vjust = 0.5),
        aspect.ratio = 5) +
  scale_color_manual(values = clrs$tumor_supersite)

3.5 sub cluster CD8 cells

# seu_obj_cd8 <- seu_obj_sub %>%
#   subset(subset = cluster_label == "CD8.T") %>%
#   FindNeighbors(reduction = "harmony", dims = 1:50) %>%
#   FindClusters(resolution = 0.2) %>%
#   RunUMAP(dims = 1:50, reduction = "harmony", reduction.name = "umapharmony", reduction.key = "umapharmony_")
# 
# write_rds(seu_obj_cd8, "/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/CD8.T_processed.rds")
seu_obj_cd8 <- read_rds("/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/CD8.T_processed.rds")

# marker_tbl <- FindAllMarkers(seu_obj_cd8, only.pos = T)
# write_tsv(as_tibble(marker_tbl), "/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/CD8.T_markers.tsv")
marker_tbl <- read_tsv("/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/CD8.T_markers.tsv")

## Hypergeometric test --------------------------------------

test_set <- marker_tbl %>% 
  group_by(cluster) %>% 
  filter(!str_detect(gene, "^RPS|^RPL")) %>% 
  slice(1:50) %>% 
  mutate(k = length(cluster)) %>% 
  ungroup %>%
  select(cluster, gene, k) %>% 
  mutate(join_helper = 1) %>% 
  group_by(cluster, join_helper, k) %>%
  nest(test_set = gene)

markers_doub_tbl <- markers_v6 %>% 
  enframe("subtype", "gene") %>% 
  filter(!(subtype %in% unique(c(coi, cell_type_major)))) %>% 
  unnest(gene) %>% 
  group_by(gene) %>% 
  filter(length(gene) == 1) %>% 
  mutate(subtype = paste0("doublet.", subtype)) %>% 
  bind_rows(tibble(subtype = "Mito.high", gene = grep("^MT-", rownames(seu_obj), value = T)))

ref_set <- markers_v6_super[["CD8.T"]] %>% 
  bind_rows(markers_doub_tbl) %>% 
  group_by(subtype) %>% 
  mutate(m = length(gene),
         n = length(rownames(seu_obj))-m,
         join_helper = 1) %>% 
  group_by(subtype, m, n, join_helper) %>%
  nest(ref_set = gene)

hyper_tbl <- test_set %>% 
  left_join(ref_set, by = "join_helper") %>% 
  group_by(cluster, subtype, m, n, k) %>%
  do(q = length(intersect(unlist(.$ref_set), unlist(.$test_set)))) %>%
  mutate(pval = 1-phyper(q = q, m = m, n = n, k = k)) %>%
  ungroup %>%
  mutate(qval = p.adjust(pval, "BH"),
         sig = qval < 0.01)

# hyper_tbl %>% 
#   group_by(subtype) %>% 
#   filter(any(qval < 0.01)) %>%
#   ggplot(aes(subtype, -log10(qval), fill = sig)) +
#   geom_bar(stat = "identity") +
#   facet_wrap(~cluster) +
#   coord_flip()
  
low_rank <- str_detect(unique(hyper_tbl$subtype), "Mito|doublet")
subtype_lvl <- c(sort(unique(hyper_tbl$subtype)[!low_rank]), sort(unique(hyper_tbl$subtype)[low_rank]))
  
cluster_label_tbl <- hyper_tbl %>% 
  mutate(subtype = ordered(subtype, levels = subtype_lvl)) %>% 
  arrange(qval, subtype) %>%
  group_by(cluster) %>% 
  slice(1) %>% 
  mutate(subtype = ifelse(sig, as.character(subtype), paste0("unknown_", cluster))) %>% 
  select(cluster, cluster_label = subtype) %>% 
  ungroup %>% 
  mutate(cluster_label = make.unique(cluster_label, sep = "_"))

marker_sheet <- marker_tbl %>% 
  left_join(cluster_label_tbl, by = "cluster") %>% 
  group_by(cluster_label) %>% 
  filter(!str_detect(gene, "^RPS|^RPL")) %>% 
  slice(1:50) %>% 
  mutate(rank = row_number()) %>% 
  select(cluster_label, gene, rank) %>% 
  spread(cluster_label, gene) %>% 
  mutate_all(.funs = helper_f)

formattable::formattable(marker_sheet)

rank	CD8.T.activated.CD40LG	CD8.T.activated.HSP	CD8.T.dysfunctional.CXCL13	CD8.T.dysfunctional.ISG	CD8.T.naive	doublet.Plasma.cell	gd.T.cell
1	CD40LG	HSPA1A	CXCL13	IFIT3	IL7R	IGHG1	TRDC
2	CD4	HSPA1B	GZMB	IFIT1	KLRG1	IGLC3	GNLY
3	IL7R	MT1X	CCL3	IFIT2	GZMK	IGKC	KLRC2
4	TNFRSF4	MT1E	HAVCR2	ISG15	GPR183	IGLC2	XCL1
5	LEF1	HSPA6	TNFRSF9	MX1	TCF7	IGHG3	FCER1G
6	TNFRSF25	DNAJB1	KRT86	RSAD2	EEF1B2	IGHGP	AREG
7	CTSL	MT1G	CTLA4	IFI6	SORL1	IGHG4	TYROBP
8	CCR7	HSPH1	PHLDA1	MX2	PLEK	JCHAIN	TRGC1
9	TIMP1	HSP90AA1	MIR155HG	HERC5	CCR7	IGHM	HOPX
10	RORA	HSP90AB1	IFNG	IFI44L	MYBL1	IGHA1	IKZF2
11	SELL	DNAJA4	CCL4L2	OAS1	RIPOR2	MZB1	NCR3
12	MAF	DNAJB4	ENTPD1	USP18	NCF1	CCL3L1	TRGC2
13	LTB	ZFAND2A	SRGAP3	OAS3	PTGDR		CD7
14	KLRB1	HSPB1	SPRY1	ISG20	PLCG2		IFITM3
15	NFKBIA	DUSP1	TNFRSF18	CMPK2			TNFRSF18
16	PLCG2	SLC30A1	AKAP5	TNFSF10			XCL2
17	TNFAIP3	FOS	VCAM1	EPSTI1			KIR3DL2
18		JUN	CLNK	IFI44			ZNF683
19		FKBP4	CD63	SAMD9L			ITGAX
20		FOSB	LAYN	NT5C3A			TXK
21		MT2A	TIGIT	IFI27			KLRD1
22		JUNB	NDFIP2	CD38			CXXC5
23		NR4A1	MYO7A	DDX58			LAT2
24		TSPYL2	FAM3C	GBP1			KIR2DL3
25		RGS2	FABP5	IRF7			LEF1
26		MT1F	PDCD1	IFIH1			CTSW
27		ERN1	HLA-DRA	LY6E			SPINT2
28		UBC	ADGRG1	OASL			SPRY2
29		ANXA1	ZNF683	DDX60L			ITGAM
30		GADD45B	CCND2	SAT1			CD63
31		IL7R	TNFSF4	OAS2			FOS
32		SLC2A3	GAPDH	LGALS9			LYN
33		PLIN2	SAMSN1	ZBP1			RRAS2
34		DDIT3	TNFRSF1B	TYMP			PRMT9
35		IER5	HLA-DQA1	MT2A			CD9
36		LMNA	LGALS3	GBP4			TRG-AS1
37		NFKBIA	HLA-DRB1	HSH2D			MATK
38		PFKFB3	CSF1	LAMP3			CD300A
39		TNF	DGKH	LAG3			ZNF331
40		NABP1	LAG3	RTP4			FCRL6
41		CREM	ACP5	GPR155			CYSLTR1
42		IER5L	DUSP4	ENDOD1			LTB
43		DDIT4	TNS3	IFITM3			IL2RB
44		ATF3	TBC1D4	ETV7			PTGDR
45		ZNF331	CCL3L1	EHD4			SRGAP3
46		HMGB2	LRRN3	SP140			PLAC8
47		AREG	CHN1	CASP1			CCDC50
48		SLC5A3	ANKRD28	H3F3B			RIN3
49			JAML	ECE1			JUNB
50			CXCR6	GBP5			PTGER2

write_tsv(marker_sheet, paste0("/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/CD8.T.cell_marker_sheet.tsv"))

seu_obj_cd8$cluster_label <- unname(deframe(cluster_label_tbl)[as.character(unlist(seu_obj_cd8[[paste0("RNA_snn_res.", louvain_resolution)]]))])

seu_obj_sub_sub <- seu_obj_sub

cluster_label_tbl_1 <- as_tibble(cbind(cell_id=colnames(seu_obj_sub), FetchData(seu_obj_sub, c("cluster_label"))))
cluster_label_tbl_2 <- as_tibble(cbind(cell_id=colnames(seu_obj_cd8), FetchData(seu_obj_cd8, c("cluster_label"))))

cluster_label_tbl <- left_join(cluster_label_tbl_1, cluster_label_tbl_2, by = "cell_id") %>%
  mutate(cluster_label = ifelse(is.na(cluster_label.y), cluster_label.x, cluster_label.y))

seu_obj_sub_sub$cluster_label <- cluster_label_tbl$cluster_label

seu_obj_sub_sub <- subset(seu_obj_sub_sub, subset = cluster_label != "doublet.Plasma.cell")
write_rds(seu_obj_sub_sub, "/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/T.cell_processed_filtered_sub.rds")

root_cell <- "SPECTRUM-OV-036_S1_CD45P_PELVIC_PERITONEUM_CGGACACCAACGACTT"
seu_obj_cd8_sub <- subset(seu_obj_cd8, subset = cluster_label != "doublet.Plasma.cell" & cluster_label != "gd.T.cell")
seu_obj_cd8_sub <- subset(seu_obj_cd8_sub, cells = c(root_cell, colnames(seu_obj_cd8_sub)[colnames(seu_obj_cd8_sub)!=root_cell][-1]))
write_rds(seu_obj_sub_sub, "/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/T.cell_processed_filtered_sub.rds")
seu_obj_sub_sub <- read_rds("/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/T.cell_processed_filtered_sub.rds")

dc_obj <- DiffusionMap(seu_obj_cd8_sub@reductions$harmony@cell.embeddings, k = 100)
dc_mat <- dc_obj@eigenvectors
colnames(dc_mat) <- paste0("DC_", 1:ncol(dc_mat))
seu_obj_cd8_sub[["DC"]] <- CreateDimReducObject(embeddings = dc_mat, key = "DC_", assay = DefaultAssay(seu_obj_cd8_sub))

write_rds(seu_obj_cd8_sub, "/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/CD8.T_processed_filtered.rds")

dpt_obj <- DPT(dc_obj)
write_rds(dpt_obj, "/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/CD8.T_sub_dpt.rds")
seu_obj_cd8_sub$DPT1 <- dpt_obj$DPT1

write_rds(seu_obj_cd8_sub, "/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/CD8.T_processed_filtered.rds")
seu_obj_cd8_sub <- read_rds("/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/CD8.T_processed_filtered.rds")

3.5.1 UMAP

plot_data_sub_sub <- as_tibble(FetchData(seu_obj_sub_sub, c(myfeatures, "cluster_label"))) %>% 
  left_join(meta_tbl, by = "sample") %>% 
  mutate(patient_id = str_remove_all(patient_id, "SPECTRUM-OV-"),
         tumor_supersite = ordered(tumor_supersite, levels = names(clrs$tumor_supersite))) %>% 
  mutate(cell_id = colnames(seu_obj_sub_sub),
         cluster_label = ordered(cluster_label, levels = my_subtypes),
         )
  
if (cell_sort == "CD45+") {
plot_data_sub_sub <- filter(plot_data_sub_sub, sort_parameters != "singlet, live, CD45-", !is.na(tumor_supersite))
}

if (cell_sort == "CD45-") {
plot_data_sub_sub <- filter(plot_data_sub_sub, sort_parameters != "singlet, live, CD45+", !is.na(tumor_supersite))
}

ggplot(plot_data_sub_sub, aes(umapharmony_1, umapharmony_2, color = cluster_label)) + 
  common_layers_disc +
  ggtitle("Sub cluster") +
  #facet_wrap(~cluster_label) +
  scale_color_manual(values = clrs$cluster_label[[coi]])

ggplot(plot_data_sub_sub, aes(umapharmony_1, umapharmony_2, color = patient_id_short)) + 
  common_layers_disc +
  ggtitle("Patient") +
  # facet_wrap(~therapy) +
  scale_color_manual(values = clrs$patient_id_short)

ggplot(plot_data_sub_sub, aes(umapharmony_1, umapharmony_2, color = tumor_supersite)) + 
  # geom_point(aes(umapharmony_1, umapharmony_2), 
  #            color = "grey90", size = 0.01, 
  #            data = select(plot_data_sub_sub, -tumor_supersite)) +
  common_layers_disc +
  ggtitle("Site") +
  # facet_wrap(~therapy) +
  scale_color_manual(values = clrs$tumor_supersite)

write_tsv(select(plot_data_sub_sub, cell_id, everything(), -umapharmony_1, -umapharmony_2, -contains("RNA_")), paste0("/work/shah/uhlitzf/data/SPECTRUM/freeze/v6/", coi, "_embedding_sub.tsv"))

3.5.2 UMAP & DC

plot_data_cd8_sub <- as_tibble(FetchData(seu_obj_cd8_sub, c(myfeatures, "cluster_label", "DC_1", "DC_2"))) %>% 
  left_join(meta_tbl, by = "sample") %>% 
  mutate(patient_id = str_remove_all(patient_id, "SPECTRUM-OV-"),
         tumor_supersite = ordered(tumor_supersite, levels = names(clrs$tumor_supersite))) %>% 
  mutate(cell_id = colnames(seu_obj_cd8_sub),
         cluster_label = ordered(cluster_label, levels = my_subtypes),
         )
  
if (cell_sort == "CD45+") {
plot_data_cd8_sub <- filter(plot_data_cd8_sub, sort_parameters != "singlet, live, CD45-", !is.na(tumor_supersite))
}

if (cell_sort == "CD45-") {
plot_data_cd8_sub <- filter(plot_data_cd8_sub, sort_parameters != "singlet, live, CD45+", !is.na(tumor_supersite))
}

ggplot(plot_data_cd8_sub, aes(umapharmony_1, umapharmony_2, color = cluster_label)) + 
  common_layers_disc +
  ggtitle("Sub cluster") +
  #facet_wrap(~cluster_label) +
  scale_color_manual(values = clrs$cluster_label[[coi]])

ggplot(plot_data_cd8_sub, aes(umapharmony_1, umapharmony_2, color = patient_id_short)) + 
  common_layers_disc +
  ggtitle("Patient") +
  # facet_wrap(~therapy) +
  scale_color_manual(values = clrs$patient_id_short)

ggplot(plot_data_cd8_sub, aes(umapharmony_1, umapharmony_2, color = tumor_supersite)) + 
  common_layers_disc +
  ggtitle("Site") +
  # facet_wrap(~therapy) +
  scale_color_manual(values = clrs$tumor_supersite)

ggplot(plot_data_cd8_sub, aes(DC_1, DC_2, color = cluster_label)) + 
  common_layers_disc +
  ggtitle("Sub cluster") +
  #facet_wrap(~cluster_label) +
  scale_color_manual(values = clrs$cluster_label[[coi]])

ggplot(plot_data_cd8_sub, aes(DC_1, DC_2, color = patient_id_short)) + 
  common_layers_disc +
  ggtitle("Patient") +
  # facet_wrap(~therapy) +
  scale_color_manual(values = clrs$patient_id_short)

ggplot(plot_data_cd8_sub, aes(DC_1, DC_2, color = tumor_supersite)) + 
  common_layers_disc +
  ggtitle("Site") +
  # facet_wrap(~therapy) +
  scale_color_manual(values = clrs$tumor_supersite)

4 session info

devtools::session_info()

## ─ Session info ───────────────────────────────────────────────────────────────
##  setting  value                       
##  version  R version 3.6.2 (2019-12-12)
##  os       Debian GNU/Linux 10 (buster)
##  system   x86_64, linux-gnu           
##  ui       X11                         
##  language (EN)                        
##  collate  en_US.UTF-8                 
##  ctype    en_US.UTF-8                 
##  tz       Etc/UTC                     
##  date     2020-12-22                  
## 
## ─ Packages ───────────────────────────────────────────────────────────────────
##  package              * version    date       lib
##  abind                  1.4-5      2016-07-21 [2]
##  ape                    5.3        2019-03-17 [2]
##  assertthat             0.2.1      2019-03-21 [2]
##  backports              1.1.10     2020-09-15 [1]
##  bibtex                 0.4.2.2    2020-01-02 [2]
##  Biobase                2.46.0     2019-10-29 [2]
##  BiocGenerics           0.32.0     2019-10-29 [2]
##  BiocParallel           1.20.1     2019-12-21 [2]
##  bitops                 1.0-6      2013-08-17 [2]
##  boot                   1.3-24     2019-12-20 [3]
##  broom                  0.7.2      2020-10-20 [1]
##  callr                  3.4.2      2020-02-12 [1]
##  car                    3.0-8      2020-05-21 [1]
##  carData                3.0-4      2020-05-22 [1]
##  caTools                1.17.1.4   2020-01-13 [2]
##  cellranger             1.1.0      2016-07-27 [2]
##  class                  7.3-15     2019-01-01 [3]
##  cli                    2.0.2      2020-02-28 [1]
##  cluster                2.1.0      2019-06-19 [3]
##  codetools              0.2-16     2018-12-24 [3]
##  colorblindr          * 0.1.0      2020-01-13 [2]
##  colorspace           * 1.4-2      2019-12-29 [2]
##  cowplot              * 1.0.0      2019-07-11 [2]
##  crayon                 1.3.4      2017-09-16 [1]
##  curl                   4.3        2019-12-02 [2]
##  data.table             1.12.8     2019-12-09 [2]
##  DBI                    1.1.0      2019-12-15 [2]
##  dbplyr                 2.0.0      2020-11-03 [1]
##  DelayedArray           0.12.2     2020-01-06 [2]
##  DEoptimR               1.0-8      2016-11-19 [1]
##  desc                   1.2.0      2018-05-01 [2]
##  destiny              * 3.0.1      2020-01-16 [1]
##  devtools               2.2.1      2019-09-24 [2]
##  digest                 0.6.25     2020-02-23 [1]
##  dplyr                * 1.0.2      2020-08-18 [1]
##  e1071                  1.7-3      2019-11-26 [1]
##  ellipsis               0.3.1      2020-05-15 [1]
##  evaluate               0.14       2019-05-28 [2]
##  fansi                  0.4.1      2020-01-08 [2]
##  farver                 2.0.3      2020-01-16 [1]
##  fitdistrplus           1.0-14     2019-01-23 [2]
##  forcats              * 0.5.0      2020-03-01 [1]
##  foreign                0.8-74     2019-12-26 [3]
##  formattable            0.2.0.1    2016-08-05 [1]
##  fs                     1.5.0      2020-07-31 [1]
##  future                 1.15.1     2019-11-25 [2]
##  future.apply           1.4.0      2020-01-07 [2]
##  gbRd                   0.4-11     2012-10-01 [2]
##  gdata                  2.18.0     2017-06-06 [2]
##  generics               0.0.2      2018-11-29 [2]
##  GenomeInfoDb           1.22.0     2019-10-29 [2]
##  GenomeInfoDbData       1.2.2      2020-01-14 [2]
##  GenomicRanges          1.38.0     2019-10-29 [2]
##  ggplot.multistats      1.0.0      2019-10-28 [1]
##  ggplot2              * 3.3.2      2020-06-19 [1]
##  ggrepel                0.8.1      2019-05-07 [2]
##  ggridges               0.5.2      2020-01-12 [2]
##  ggthemes               4.2.0      2019-05-13 [1]
##  globals                0.12.5     2019-12-07 [2]
##  glue                   1.3.2      2020-03-12 [1]
##  gplots                 3.0.1.2    2020-01-11 [2]
##  gridExtra              2.3        2017-09-09 [2]
##  gtable                 0.3.0      2019-03-25 [2]
##  gtools                 3.8.1      2018-06-26 [2]
##  haven                  2.3.1      2020-06-01 [1]
##  hexbin                 1.28.0     2019-11-11 [2]
##  hms                    0.5.3      2020-01-08 [1]
##  htmltools              0.4.0      2019-10-04 [2]
##  htmlwidgets            1.5.1      2019-10-08 [2]
##  httr                   1.4.2      2020-07-20 [1]
##  ica                    1.0-2      2018-05-24 [2]
##  igraph                 1.2.5      2020-03-19 [1]
##  IRanges                2.20.2     2020-01-13 [2]
##  irlba                  2.3.3      2019-02-05 [2]
##  jsonlite               1.7.1      2020-09-07 [1]
##  KernSmooth             2.23-16    2019-10-15 [3]
##  knitr                  1.26       2019-11-12 [2]
##  knn.covertree          1.0        2019-10-28 [1]
##  labeling               0.3        2014-08-23 [2]
##  laeken                 0.5.1      2020-02-05 [1]
##  lattice                0.20-38    2018-11-04 [3]
##  lazyeval               0.2.2      2019-03-15 [2]
##  leiden                 0.3.1      2019-07-23 [2]
##  lifecycle              0.2.0      2020-03-06 [1]
##  listenv                0.8.0      2019-12-05 [2]
##  lmtest                 0.9-37     2019-04-30 [2]
##  lsei                   1.2-0      2017-10-23 [2]
##  lubridate              1.7.9.2    2020-11-13 [1]
##  magrittr             * 2.0.1      2020-11-17 [1]
##  MASS                   7.3-51.5   2019-12-20 [3]
##  Matrix                 1.2-18     2019-11-27 [3]
##  matrixStats            0.56.0     2020-03-13 [1]
##  memoise                1.1.0      2017-04-21 [2]
##  metap                  1.2        2019-12-08 [2]
##  mnormt                 1.5-5      2016-10-15 [2]
##  modelr                 0.1.8      2020-05-19 [1]
##  multcomp               1.4-12     2020-01-10 [2]
##  multtest               2.42.0     2019-10-29 [2]
##  munsell                0.5.0      2018-06-12 [2]
##  mutoss                 0.1-12     2017-12-04 [2]
##  mvtnorm                1.0-12     2020-01-09 [2]
##  nlme                   3.1-143    2019-12-10 [3]
##  nnet                   7.3-12     2016-02-02 [3]
##  npsurv                 0.4-0      2017-10-14 [2]
##  numDeriv               2016.8-1.1 2019-06-06 [2]
##  openxlsx               4.1.5      2020-05-06 [1]
##  pbapply                1.4-2      2019-08-31 [2]
##  pcaMethods             1.78.0     2019-10-29 [2]
##  pillar                 1.4.6      2020-07-10 [1]
##  pkgbuild               1.0.6      2019-10-09 [2]
##  pkgconfig              2.0.3      2019-09-22 [1]
##  pkgload                1.0.2      2018-10-29 [2]
##  plotly                 4.9.1      2019-11-07 [2]
##  plotrix                3.7-7      2019-12-05 [2]
##  plyr                   1.8.5      2019-12-10 [2]
##  png                    0.1-7      2013-12-03 [2]
##  prettyunits            1.1.1      2020-01-24 [1]
##  processx               3.4.2      2020-02-09 [1]
##  progeny              * 1.11.3     2020-10-22 [1]
##  proxy                  0.4-24     2020-04-25 [1]
##  ps                     1.3.2      2020-02-13 [1]
##  purrr                * 0.3.4      2020-04-17 [1]
##  R.methodsS3            1.7.1      2016-02-16 [2]
##  R.oo                   1.23.0     2019-11-03 [2]
##  R.utils                2.9.2      2019-12-08 [2]
##  R6                     2.4.1      2019-11-12 [1]
##  ranger                 0.12.1     2020-01-10 [1]
##  RANN                   2.6.1      2019-01-08 [2]
##  rappdirs               0.3.1      2016-03-28 [2]
##  RColorBrewer           1.1-2      2014-12-07 [2]
##  Rcpp                   1.0.4      2020-03-17 [1]
##  RcppAnnoy              0.0.16     2020-03-08 [1]
##  RcppEigen              0.3.3.7.0  2019-11-16 [2]
##  RcppHNSW               0.2.0      2019-09-20 [2]
##  RcppParallel           4.4.4      2019-09-27 [2]
##  RCurl                  1.98-1.1   2020-01-19 [1]
##  Rdpack                 0.11-1     2019-12-14 [2]
##  readr                * 1.4.0      2020-10-05 [1]
##  readxl               * 1.3.1      2019-03-13 [2]
##  rematch                1.0.1      2016-04-21 [2]
##  remotes                2.1.0      2019-06-24 [2]
##  reprex                 0.3.0      2019-05-16 [2]
##  reshape2               1.4.3      2017-12-11 [2]
##  reticulate             1.14       2019-12-17 [2]
##  rio                    0.5.16     2018-11-26 [1]
##  rlang                  0.4.8      2020-10-08 [1]
##  rmarkdown              2.0        2019-12-12 [2]
##  robustbase             0.93-6     2020-03-23 [1]
##  ROCR                   1.0-7      2015-03-26 [2]
##  rprojroot              1.3-2      2018-01-03 [2]
##  RSpectra               0.16-0     2019-12-01 [2]
##  rstudioapi             0.11       2020-02-07 [1]
##  rsvd                   1.0.3      2020-02-17 [1]
##  Rtsne                  0.15       2018-11-10 [2]
##  rvest                  0.3.6      2020-07-25 [1]
##  S4Vectors              0.24.2     2020-01-13 [2]
##  sandwich               2.5-1      2019-04-06 [2]
##  scales                 1.1.0      2019-11-18 [2]
##  scatterplot3d          0.3-41     2018-03-14 [1]
##  sctransform            0.2.1      2019-12-17 [2]
##  SDMTools               1.1-221.2  2019-11-30 [2]
##  sessioninfo            1.1.1      2018-11-05 [2]
##  Seurat               * 3.1.2      2019-12-12 [2]
##  SingleCellExperiment   1.8.0      2019-10-29 [2]
##  smoother               1.1        2015-04-16 [1]
##  sn                     1.5-4      2019-05-14 [2]
##  sp                     1.4-2      2020-05-20 [1]
##  stringi                1.5.3      2020-09-09 [1]
##  stringr              * 1.4.0      2019-02-10 [1]
##  SummarizedExperiment   1.16.1     2019-12-19 [2]
##  survival               3.1-8      2019-12-03 [3]
##  testthat               2.3.2      2020-03-02 [1]
##  TFisher                0.2.0      2018-03-21 [2]
##  TH.data                1.0-10     2019-01-21 [2]
##  tibble               * 3.0.4      2020-10-12 [1]
##  tidyr                * 1.1.2      2020-08-27 [1]
##  tidyselect             1.1.0      2020-05-11 [1]
##  tidyverse            * 1.3.0      2019-11-21 [2]
##  tsne                   0.1-3      2016-07-15 [2]
##  TTR                    0.23-6     2019-12-15 [1]
##  usethis                1.5.1      2019-07-04 [2]
##  uwot                   0.1.5      2019-12-04 [2]
##  vcd                    1.4-7      2020-04-02 [1]
##  vctrs                  0.3.5      2020-11-17 [1]
##  VIM                    6.0.0      2020-05-08 [1]
##  viridis              * 0.5.1      2018-03-29 [2]
##  viridisLite          * 0.3.0      2018-02-01 [2]
##  withr                  2.3.0      2020-09-22 [1]
##  xfun                   0.12       2020-01-13 [2]
##  xml2                   1.3.2      2020-04-23 [1]
##  xts                    0.12-0     2020-01-19 [1]
##  XVector                0.26.0     2019-10-29 [2]
##  yaml                   2.2.1      2020-02-01 [1]
##  zip                    2.0.4      2019-09-01 [1]
##  zlibbioc               1.32.0     2019-10-29 [2]
##  zoo                    1.8-7      2020-01-10 [2]
##  source                                 
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.3)                         
##  CRAN (R 3.6.2)                         
##  Bioconductor                           
##  Bioconductor                           
##  Bioconductor                           
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  Github (clauswilke/colorblindr@1ac3d4d)
##  R-Forge (R 3.6.2)                      
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.3)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  Bioconductor                           
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  Bioconductor                           
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.3)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.3)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  Bioconductor                           
##  Bioconductor                           
##  Bioconductor                           
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.3)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  Bioconductor                           
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  Bioconductor                           
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  Bioconductor                           
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.3)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  Github (saezlab/progeny@94bea1c)       
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.3)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.3)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  Bioconductor                           
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  Bioconductor                           
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.3)                         
##  CRAN (R 3.6.3)                         
##  Bioconductor                           
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.3)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  Bioconductor                           
##  CRAN (R 3.6.2)                         
##  CRAN (R 3.6.2)                         
##  Bioconductor                           
##  CRAN (R 3.6.2)                         
## 
## [1] /home/uhlitzf/R/lib
## [2] /usr/local/lib/R/site-library
## [3] /usr/local/lib/R/library

MSK SPECTRUM freeze major subset deep dives

Florian Uhlitz

2020-12-22

1 Cluster markers

1.1 Major T.super markers for cell assign

1.2 Subtype currated markers

1.3 Subtype cluster markers

2 Clusters

2.1 sizes

2.2 UMAP

3 Filtering out doublet and mito high clusters

3.1 UMAP

3.2 add module scores and pathway scores

3.3 marker heatmap

3.4 composition

3.4.1 per site

3.4.2 per sample

3.4.3 site specific cluster enrichment

3.5 sub cluster CD8 cells

3.5.1 UMAP

3.5.2 UMAP & DC

4 session info